Data-driven decision-making of marine ecological civilization construction in island county of China

In the cause of assessing the construction of ecological civilization in island county accurately, an evaluation index system involving 27 single indexes under the framework of Pressure-State-Response (PSR) model is established. Because the system of island county ecological civilization development is a complex network system. This research designs an assessment method based on network analytical method to evaluate the development of island county ecological civilization. For this purpose, a Back-Propagation (BP) neural network-based decision-making model are established at the same time. Regarding the standardized data of 6 island counties from 2010 to 2020 in China as the input learning matrix, both entropy and coefficient of variation are applied to the evaluation of the “Development Index of Marine Ecological Civilization Construction” (DIMECC). More specifically, after 10,000 times of training, the error accuracy is within overhead manageable, and the island counties which are evaluated as ecological civilization level are found out and the key impacting indicator contributing to low level are identified. According to the actual evaluation result of island counties obtained by BP neural network, the ecological civilization of them is proposed as a whole under land-sea coordination. Finally, this paper suggests on the decision-making of the participants to contribute the construction of the island county ecological civilization development system.